Department of Biostatistics Seminar/Workshop Series
A New Linear Model-Based Approach for Modeling the Area Under the Outcome-Time Curve (AUC)
Rameela Chandrasekhar
Department of Biostatistics, University at Buffalo
Friday, March 18 11:00am-12:00pm, Vanderbilt Children's Theater
Outcome versus time data is commonly encountered in biomedical and clinical research. A common strategy adopted in analyzing such longitudinal data is to condense the repeated measurements on any individual to a single summary statistic. Scientists find the use of summary statistics appealing since it offers an integrated
approach in representing the patient's overall response and is also easy to interpret. A commonly used summary statistic in such studies for comparison between groups is the area under the curve (AUC), which is commonly calculated using the trapezoidal rule. Standard parametric (t-test) or non-parametric methods (wilcoxon rank
sum test) are then applied to test for differences in the AUC distribution. Disadvantages of this approach include the disregard of the within-subject variation in the longitudinal profile. We propose a general linear model approach accounting for the within-subject variance for estimation and hypothesis tests about the mean areas. Inferential properties of our approach such as Type I error and power are compared to those from standard methods of analysis using simulation studies. The impact of missing data, within-subject heterogeneity and homogeneity of within-subject variance are also evaluated. This approach is extended to the assessment of bioequivalence and its properties are investigated.